Ben Horowitz 与 David Solomon：40年来最甜蜜的宏观窗口

We were the largest wholesale funder in the world 10 years ago. There are a lot of things you want to be the largest in the world. Wholesale funding is not one of them.

We got a lot of criticism: “Why are you raising money now? What, are you stupid?” It turns out that the best time to raise money is when nobody has money.

Last year, the 4 largest companies contributed 1% to GDP growth with their $400 billion of spending.

David, you've been at Goldman now for over 25 years. What are you focused on to position Goldman for the future?

If you're in our kind of businesses, if you're attached to financial assets, this is as sweet a spot as I've seen.

With AI, if you have proprietary data and enough GPUs, you can solve almost any problem. It is magic.

I've had the distinct pleasure of working, at least indirectly, for both David Solomon and Ben Horowitz, and I have a lot of affection for both Goldman Sachs and a16z. If you haven't read it, I highly recommend reading The Partnership, written by Charles Ellis, which chronicles Goldman's nearly 160-year history.

I think the most remarkable thing about Goldman's history is the fact that it's not a business built through a series of bank mergers. Unlike many of its peers, it was really a business built brick by brick by generations of entrepreneurial partners raising their hands, going off, and building new businesses, whether it was expanding into Europe, starting the merchant banking business, or creating the wealth management division. Many of these business units became global franchises.

I'd argue that Goldman was and still is one of the most entrepreneurial financial institutions in the world. As I think about where we are in our own evolution at Andreessen Horowitz, I like to think that this is what Goldman Sachs must have felt like 50 or 75 years ago: a small group of entrepreneurial investors betting on a future—

They weren't as rich as you guys, though.

Also, Goldman stopped speaking to Sachs, like, forever.

They got very mad at each other over—was it Sachs who supported Germany in World War I?

So you actually remember your history. Wow.

Yeah.

Well, yeah, a small partnership betting on the future with big hopes and ambitions. I'll leave it at that.

Well done, David.

Thank you. But maybe just pulling on that thread: David, you've been at Goldman now for over 25 years. You joined the firm, I believe, in 1999, just after the firm's IPO. How has the firm evolved during your tenure? And maybe more importantly, what are you focused on to position Goldman for the future?

First of all, it's great to be here and great to be with everybody. Before I start on that, I'd just say one of the big lessons I have in my life is: If you're joining a new firm and it's a private partnership, don't spend 6 months negotiating so that you carry over past the IPO date. Join before the IPO. It's a good lesson for all of you in private partnerships.

The firm is a remarkable place, and I really appreciate what you said about the firm's entrepreneurial spirit. The firm was a private partnership for a long time. The thing about private partnerships is that you have this mutual agency where people go off and do things. There's some structure that creates a collective each year or each cycle, where everything comes back, and then there's a reevaluation of the partnership shares. Then you go off again into the future to do more.

That served the firm incredibly well, and the firm stayed a partnership much longer than any other really big Wall Street firm. But I'd like to say that the firm stayed a partnership until the last moment when it absolutely couldn't be a partnership anymore, because it needed the permanent capital to really make it a relevant business.

If the firm hadn't gone public in 1999, it would have missed the global expansion of capital markets and probably would look more like Lazard today than like Goldman Sachs. The stewards of the firm at that point did an incredible job.

I think the challenge for us over the last 25 years—and I think the leadership team over the last 8 years has really done an incredible job working together to do this—is somehow, 25 years after an IPO, we still have this partnership culture. It's highly aspirational. Every 2 years, we have 450 people who become partners of Goldman Sachs, and they're really compensated in correlation with how the overall enterprise does.

The big thing that I'm really proud of that we, as a broad leadership, have done is that we've started to recognize that we're not a small private partnership. You can't be a public company and not grow and have some form of top-down strategic direction that really gets the whole thing working—making 1 plus 1 plus 1 plus 1 equal more than what the math adds up to. That's been a journey, and it's been bumpy. You were there for part of those bumps.

But I think we've navigated well. I still think the principles and values that we sit upon as a firm are strong. We really strive to be the most exceptional financial institution in the world. We don't always get there, but we strive for that. We really sit on 4 core values: client service, partnership, integrity, and excellence. We try to live it, and I think the firm's in a really good place.

In some ways, it hasn't changed at all in 26 years. In some ways, it's changed massively.

Are there a few things you're most focused on as CEO, looking forward over the next 5 or 10 years?

One of the things I've learned is that I was a banker and advised CEOs for a lot of my career, but actually owning the responsibility is very different from giving advice. I think the most important thing that a CEO has to do in a big enterprise like this is own the strategy and the direction of the firm.

I'm focused on how we ensure we're executing toward growing the firm, because I know we have to do that to perform on a relative basis. But then I'm also thinking about and worrying about big-picture strategic risks that can make the firm less relevant, less successful, less important, or less competitive.

For us, I think there are 2 things that the firm is really focused on. First of all, one of the things that makes the United States an extraordinary place is that we have the most extraordinary capital markets, the most extraordinary financial system, and the most extraordinary financial institutions.

I would argue that the 6 most important financial institutions in the US are all US financial institutions, and there is no global institution that can compete in terms of its relevance in the world with the 6 most important US institutions. When you look at those institutions, there are different kinds.

There are retail banks—more traditional, banky banks. That would include JPMorgan, Wells Fargo, Bank of America, and Citibank. They all have global businesses, but they are truly banks in what they do. They have retail platforms and retail businesses.

Then you've got 2 institutional firms. That doesn't mean they don't touch individuals in different ways, but Morgan Stanley and Goldman Sachs are both institutional firms. Goldman Sachs is a little bit of an island of 1 in the context of the way we're positioned as an institutional firm, and Morgan Stanley is a little bit of an island of 1 in terms of the way they're positioned.

Scale matters a lot.

I just went through all those firms. The 2 smallest firms of all those firms are Goldman Sachs and Morgan Stanley. When there's turbulence in the world, you always want scale. Scale in these businesses, because they're so mature, gives you enormous leverage and latitude.

We continue to think a lot about scale, and we think out 5, 10, and 15 years: How are we going to maintain a level of scale that makes us competitive? Ten years ago, it would have been unfathomable that Goldman Sachs could have a $1.9 trillion balance sheet. At the moment, JPMorgan has a $4.5 trillion balance sheet. When JPMorgan's at $6 trillion, we're going to have to be at least $3.5 trillion.

We have to think about how we can continue to create that scale, because these are very mature businesses, and it's hard to really build that scale purely organically. So, that's one.

Two, funding. Funding these enterprises is one of the big strategic risks to these enterprises. These enterprises live on funding and liquidity, and we don't have a traditional deposit-funding platform.

We've got—and you participated in this—a very excellent digital deposit platform that now has over $200 billion in deposits. We've also—we have about $500 billion of total deposits. Fifteen years ago, we had zero.

So we fund about 40% of the firm with deposits, but deposits are a much more stable funding source than institutional wholesale funding.

Commercial paper.

Yeah. We were the largest wholesale funder in the world 10 years ago. There are a lot of things you want to be the largest in the world. Wholesale funding is not one of them.

So, strategically, that's another thing we're about. Those are big things: stepping back from the day-to-day execution and thinking 10, 15, 20 years out. By the way, I won't be here running the firm, but it's still my responsibility to steward and chart that.

In the short term, I'm much more focused on technology across the organization: how technology shifts the way we do things, how we're rebuilding processes, and how we're operating differently while staying true to what we do.

Awesome. Well, we're here to help with that today, too.

Absolutely.

Ben, maybe transitioning to you: you and Marc started the firm at an auspicious time, in the wake of the financial crisis, in 2009.

2009.

It turned out to be a really interesting moment because it was the beginning of mobile and the rise of the cloud.

Well, it's funny, too: we got a lot of criticism in venture capital. People asked, “Why are you raising money now? Are you stupid?” It turns out that the best time to raise money is when nobody has money. It's very obvious when you say it that way, but the nature of investing is that people always want to invest high, and they always want to walk away when the market is low. So we got very fortunate, I think.

Maybe you can describe the evolution of the firm since you started and, again, what your ambitions are for the future as well.

The original idea in venture capital is that the fundamental thing you have to be is what's known as top-tier, because if you're not top-tier, the best entrepreneurs won't take your money. There are times when the market is so blazing hot that you can be a not-important venture capital firm, dump into good deals, and make money. But most of the time, if you're not top-tier, you're going to go out of business. So you have to be that.

The difficult thing about being top-tier is that, historically, the way you became top-tier was reputationally. If you're Sequoia, you had invested in Apple, Cisco, Yahoo, and Google. It's really hard to make up that ground if you're starting in 2009.

The idea we originally had to get to top-tier was to have a better product, specifically a better product for entrepreneurs. The venture capital product was great for LPs, but we thought it was mediocre for entrepreneurs. So we designed the firm to really enable a founder to build his or her own company and run it as CEO, which wasn't really an idea then. The idea was much more to replace the founder.

Because we were founders, we knew what that was, so we created a firm to give a founder a brand, power, access, and all these kinds of things. VCs said they did those things, but they didn't have to because they were top-tier. It didn't matter. We did that, and that's how we got into position to be a long-lasting firm.

The second phase was really based on something that Marc wrote in 2011 called Why Software Is Eating the World. The idea with Why Software Is Eating the World was that, if you looked at venture capital up to that point, there were studies that said in any given year there were approximately 15 technology companies that got to $100 million in revenue. Those were going to be the companies worth money, and nothing else was going to be worth money.

The whole venture capital sport was how many of those 15 you could get into. But if software was going to eat the world, we thought maybe that 15 was going to be 150. Maybe one of the features of a venture capital firm was going to be that you had to be able to scale it.

Traditionally, I remember David Swensen—the great David Swensen, who ran the Yale endowment for years—saying to me, “A good venture capital firm is like a basketball team: 5, maybe 6 players. That's it.” But you can't address a market where you have to be in 150 companies with 6 players. How do you organize? How do you scale? How do you design the firm so that you can get to the whole opportunity and yet still be really, really good at investing, without having more than 5 or 6 people talking about a deal?

That was phase 2, and that's really when we somewhat left the building in terms of what was going on in Silicon Valley, because nobody else was thinking that way. In 2025, about 18.3% of all venture capital raised in the United States was raised by us. So we're now—from tier 1 to the biggest.

Going forward, what I think that looks like—and I get a lot of this thinking from my old mentor, Andy Grove—is important. He was actually at the end of his life, but one of the things he said to me that I always remember was profound in its obviousness. For those of you who don't know him, he ran Intel, got it through that great memory crisis, and changed the company—probably the greatest technology CEO we've seen.

He said that if you're the leader of an industry, then the growth of that industry depends on you. You have to grow the market; nobody else is going to do it. That is coming down on you. He really took that seriously at Intel.

When I think about what we are as a firm, a lot of it is incumbent on us. A lot of the work we've done on policy for crypto, the things we're doing internationally, and the things we're doing on American dynamism come down to these questions: How do we win—not just as Andreessen Horowitz, but how does the country win technologically? How do we continue to compete with China? How do we remain relevant in the next 100 years, as we were in the last 100 years?

Awesome. Maybe we'll transition just a little bit to markets. David, how would you describe the macro environment? What are you hearing from the CEOs you work with and advise most closely?

Sure. Just Ben and I were talking about this. Good times. If you're in our kind of businesses, if you're attached to financial assets or investable assets, this is—as someone who's been doing this for 40-odd years—as sweet a spot as I've seen in the macro picture.

That doesn't mean there aren't all sorts of difficult, complex things going on in the world. But let's just stay here in the United States for a minute. We can go around the world and talk about anywhere you want, but let's just start here in the United States.

The combination of the significant amount of fiscal stimulus, which is continuing to increase—the Big Beautiful Bill that started in 2026 just adds more to that—is powerful. It's not that we weren't already in a very stimulative place; we just added a whole bunch more. We have fiscal stimulus, and we have monetary stimulus because we're in a rate-cutting cycle. That doesn't mean I think we're going to see many more rate cuts, but we're probably going to see a couple more.

We are in a capital-investment supercycle, like something we've never seen. Last year, the 4 largest companies contributed 1% to GDP growth with their $400 billion of spending. We are also in a deregulatory unwind cycle, from a massive regulatory surge during the last administration to a deregulatory windback that is very stimulative.

All these things together create such a cocktail of stimulus that it's very, very hard to slow the economy down. Average Americans definitely feel a lot of stress because everything's more expensive. You could talk about inflation going from 9 to 3, but the bottom line is that everything is 25% to 30% more expensive. That's the way Americans feel it.

There's pressure, but at the same time, there's enormous financial leverage that keeps the economy going and makes the economy a little bit more versatile. If you own monetary assets or investable assets, or if you're around growth and technology, this is a pretty prime environment.

I'll give you 100 things that can set it off. Last April, if you were in Davos the previous January, people felt the same way. Then, in April, we had a speed bump, but it was only a short speed bump.

There are 2 things that I think have the market moving ahead. One, you've got a president who, if you look at the speed bump last April, marks to market to that market every single day. If the market's going in the wrong direction, he has no problem adjusting very, very quickly. Number 2 is the productivity gains from AI investment, putting it into the enterprise, and having the enterprise pick it up. The market is pulling forward a lot of what it expects to be delivered over the next 1, 2, 3, or 4 years.

And so that's a pretty prime macro environment. Now, geopolitics is much tougher. We're moving back to a multipolar world, and the risk of a geopolitical problem that really slows down growth is—I’m not saying it's high, but it's much higher than it's been for the last 10, 20, or 30 years, since the Wall fell.

The world is fragile. Social media creates a lot of volatility and division. The way people absorb information and the way information moves make the world faster-moving, but also more volatile. A lot can go wrong, but at the moment, from a base economic perspective, that cocktail of stimulus is pretty powerful.

Maybe a follow-up question for both of you. Do you expect to see a lot of M&A or IPOs this year? How are you advising your CEOs?

We have a yes.

A lot of them are in the audience. It's a good banker response—

Just fact-based. Okay, basically, we had a very, very tough regulatory environment. M&A, capital raising, and IPOs are driven by confidence, and if you have a tough regulatory environment, that affects confidence. From an M&A perspective, on strategic M&A, for the last 4 years, whatever the question was, the answer was no.

Right.

Okay, now whatever the question is, the answer—even if it's very, very significant—is maybe. So what do CEOs like to look forward to? They like to do big things. They want to do big things, and so there's a lot of activity, in fact. I just think, again, this is an environment where you're going to see significant activity. I think this could be the biggest M&A year. This is just me predicting: I think it'll be the biggest M&A year in history this year.

It's going to be a bigger IPO year, the reason being that a bunch of these big companies are finally deciding they want to come through the pipe. But you'll have a view on that too.

Being a public company is a horrible thing. I do not rant. Do not rant. [Laughter] It is challenging.

You just have to be okay with getting sued—

A lot.

All the time. You know, it's funny. We had a company that just went public, and they're like, “We might get sued.” I was like, “Of course you're going to get sued. You're public. This is America.” Like, what are you talking about?

I agree. I agree a lot on the M&A front, except that it's not clear what the FTC's position on these things is yet.

Yeah, especially on big tech. Even on small tech, they've been very, very aggressive.

So I think M&A will happen, but it may happen more in the form of IP transactions and that kind of thing than as traditional M&A. I hope not, but that may be the case. And then, yeah, look, I think there's going to be a lot of IPOs coming out of our world.

I think there's going to be some out of necessity because the companies are growing so fast. We have so many companies that went from 0 to over $100 million in less than a year, and some went from 0 to over $1 billion in less than a year. We've never seen that before.

The corollary to that in AI is that leads aren't what they once were. For my whole life in technology, and for the whole history of software, there was this thing called The Mythical Man-Month. The way The Mythical Man-Month works is: 9 women cannot have a baby in 1 month. You can't just—if you're Google—put 1,000 software engineers on a product and wipe out a startup, because you can only build that product with, say, 7 or 8 people. Once they figured it out, they've got that lead, and you're going to have to be behind for a long time.

That's not true with AI. With AI, if you have data, particularly proprietary data, and you have enough GPUs, you can solve almost any problem. It is magic. But it means that you can throw money at the problem, and we've never had that in tech.

I think that's actually going to drive a lot of IPOs, because people are going to want to get out and have the capital to continue to compete, because it's really necessary. You don't just have a lead you can sit on. So it's going to be a very exciting year, I think.

You were talking about the FTC earlier. I know you and Marc are spending a lot more time in D.C. than you ever have. What are some of the policy agendas you're most focused on, and why do you think this is more important now than it's ever been?

Well, the first one was crypto, because we thought then, and we continue to think, that crypto is an extremely important technology. It's not just the most profound breakthrough in financial technology that we've seen, but a real breakthrough in how society works.

Everything from how property rights work on the internet to what the right architecture is for things where creatives contribute most of the value, what the right business architecture is, and what stakeholder capitalism really is—these are all things that get solved with crypto. We thought it was so important for the advance of society and to not have us descend into communism and these kinds of things.

It got completely banned by the last administration, but not through a legal process or a legislative process—just through sheer will and, we'd say, abuse of the power of the government, including techniques like debanking. Our company got Wells notices, which I've never seen before in a private company. It was just an attack from the government on a technology industry in this country.

So we were like, well, we've got to get in and work on that. The first thing was the GENIUS Act, the stablecoin bill, which passed and is now law, and we're very proud of that. The second one, which we think is the more important bill, is the Clarity Act, also known as market structure.

It establishes market structure, and it's such a necessary thing for this technology because you have these tokens that can represent a Pokémon card, a stock certificate, or a dollar. There were no rules to say, well, which one is this token? The approach of the Biden administration was that everything's a security, to the point where they sued artists. It was like, “I painted a picture and I made an NFT.” “Oh, you sold a security.” That's crazy.

We're trying to get the Clarity Act passed right now. We've had some drama around it, which I'm not going to comment on, but that's a thing.

The second one that's really important is AI. With the automobile or with electricity, people freak out about new technologies because they do have a big impact. They are going to change the world. With AI in particular, some of the calls are coming from inside the house, where people are really trying to scare the population, sometimes to achieve regulatory capture and other things.

But if you ban the technology, which some people are calling for, or infringe people's ability to do mathematics, which a lot of people are calling for, then I think we're definitely going to lose the AI race to China, which has massive, 100-year implications.

The key things we're trying to protect are, first, the model is the model. It is a mathematical model. It predicts things. It's not a sentient being. Maybe we'll figure out how to do that; we don't know how to do that yet. So it's not sentient. It's just a model.

We're trying to say, don't regulate math. Regulate the applications of that math. If somebody uses AI to break into a bank, steal your money, or make a robot that shoots somebody, then that's illegal. But the technology itself shouldn't be illegal.

The most pressing issue right now is that every state wants to have its own set of AI laws, which will basically make it impossible for new companies to innovate, because you can't comply with 50 different laws from 50 different states. So we're trying to get that done shortly.

Following that, there's the issue of how copyrights are treated. Can you build a statistical model over copyrighted work—not reproduce the copyrighted work, but just build a model about it so that the software becomes smarter?

We think that's very important because China absolutely doesn't respect copyrights. They don't respect just copying it, let alone building a statistical model. We're going to have weaker AI if we can't train on the complete data.

Those are the main things that we're trying to push forward.

Awesome. One of the things that was very evident to me during my time at Goldman was how client-centric the firm was, and I know One GS was a big focus of yours. I'm curious how AI is changing the way you both work internally and how you're delivering better results for clients.

Sure. Well, the firm's business is serving our clients. Technology has, for decades and decades and decades, been making productive people more productive. Goldman Sachs is a professional services firm filled with productive people who are very productive, and technology has been changing the way they work, evolving the way they work, making them more consequential, and allowing them to expand the scope and footprint of what they impact.

This technology is another acceleration of that, for sure. In the simplest form—and this is a broad oversimplification, so please take it as such—there are 2 things that we're focused on.

One, we've got lots of smart people. These are tools and applications. We're trying to get them into their hands and give them access to them, access to models, and access to applications, so that they can experiment with them, play with them, and figure out how, on a day-to-day basis, as they're executing for clients and doing the things they're doing, they can be more productive, more powerful, and have more impact.

We're good at this. We've done this before. Our people are good at it. It takes time, but we know how to do this, and we're doing it.

It's really constrained by how we get the best tools, the best models, and the best applications—and get them, by the way, regulatory-cleared—because we have to deal with regulatory constraints in everything we touch and do. That's a huge barrier for us. We're just not a company that can say, "Oh, this is great. Let's try it." We have to have a huge process before we can try anything.

But we know how to do that. We're doing that, and that is expanding the productivity of our people. You see real-time uptake on that, and that's really accelerating.

The more interesting thing to me as the CEO is that this technology allows us to really look at fundamental operating processes in a massive enterprise and completely reimagine them—to automate them and make them more efficient. Not just simply for the benefit of doing them with fewer people or at lower cost, but for the benefit of taking some of those savings and giving us more capacity to invest in growth areas of the business where we're constrained.

We don't have the ability to just spend as much money as we want and lose as much money as we want. We actually have to be held accountable every year for how much money we spend, how much money we make, and what kind of return we generate.

Returns don't generally last forever.

No, they don't last forever. But interestingly, there are companies that have proven they can last for 10 or 15 years, where accountability for how you're deploying your capital is put off for a long period of time. We have to look at it every year.

I would say that in the last few years, we've been constrained. Last year, we spent $6 billion on technology. I would have loved to spend $8 billion, but if I spent $8 billion, our returns would have been hundreds of basis points lower. You know what? We couldn't do that.

Sure. Now, if we can actually find $2 billion of efficiency by reimagining processes, then I can spend $8 billion and wind up with the same returns. So we laid out—we actually called it One GS 3.0—a program where we picked 6 specific processes in the firm and said, "We are going to do the work to really completely reimagine them."

We have not put out publicly how that changes the workforce or how much capacity that creates, but it's super significant. It's not that there are only 6; these are just the first 6.

This is one of the reasons why the market's running forward. I think this opportunity is huge, but this is hard. This is hard because you're asking people to take away their empire and do their empire differently.

It's got to be driven top-down, and it's hard, but we're going to make a lot of progress. Those are 2 simplifications, but they're 2 big things that I like to sum up.

Anything, Ben, you'd add to that? Just where do you sort of see the proliferation of this technology in the enterprise, and what are you most optimistic about in the next, I don't know, 5 to 10 years?

Well, I think that, for the reasons David cited, we're at the very, very beginning in the enterprise. Changing people and processes and so forth in a big existing company is complicated, no matter what the technology is.

In our firm, as you know, we are taking a very aggressive approach to first automating all the things people do and don't like to do. It's not the funnest part of the job. We've also gotten all of our hard data into a Databricks data lake, so we can ask basically any question about the firm or the portfolio. Customer support for it works fantastic.

You know, agentic investing is going to be very, very interesting because models work on the facts that are available. One of the things about investing is that sometimes the biggest changes, and the way you have to think about investing in a portfolio, come from things that are completely new and unexpected. It can't be incorporated into a model. It can't be something from the past.

Yeah, it can't be something from the past. So the bottom line is, once it happens, it can be quickly incorporated. Models can load very quickly, but still, you start from a place. I'm really interested to see how it works.

And look, one of the things you've got to wonder is why there are a handful of people who have so outperformed as investors over a long period of time, but, speaking generally, you encounter people who underperform. So if the models are based on the information that all the people who are underperforming have, it's going to be interesting to see whether something different comes out of it.

Awesome. I think we're running out of time, but maybe one last bonus question: favorite DJ. No self-nominations.

Favorite DJ today?

Yeah. Or it could be in the past.

John Summit.

Okay.

I mean, John Summit is doing really, really cool things as a DJ. He's an incredibly interesting young guy who's got a lot of energy, and he's evolving very much the context of how big club-house DJs do what they do. I think he's doing a great job.

Ben, I'm going to stay in my lane, which is the past and hip-hop DJs. I'm going to say DJ Jazzy Jeff, who is very underrated because his partner, the Fresh Prince, became Will Smith. But DJ Jazzy Jeff is a great all-time DJ. Yeah. Awesome. Thank you guys so much for doing this.

Thank you, David.

Awesome.